Losing children and mental well-being: evidence from China

ABSTRACT

This article explores the impact of losing a child, especially losing all children (including losing the only child), on the mental health, happiness and loneliness of parents. The Chinese government has implemented strict restrictions on the number of births for each family since the 1970s, resulting in the creation of millions of only-child families. Using the 2011 baseline data from the China Health and Retirement Longitudinal Study, we find that the bereavement of a child is associated with lower levels of mental health and happiness and higher levels of loneliness for the parents, even after we adjust for demographic and socioeconomic characteristics. The effects are significantly stronger for losing all the children. Results have strong implications for the population-control, elderly-care and mental-care policies especially in China and developing countries, where the social security system is not yet sound.

KEYWORDS:

• Mental health
• happiness
• losing children
• one-child policy
• elderly-care policy

JEL CLASSIFICATION:

• I10
• I31
• J13
• J14

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Among those losing all their children, 86% of them are bereaved with their only child.

² The complete details of these exceptions can be found in the website of National Population and Family Planning Committee of P. R. China, http://www.chinapop.gov.cn/

³ Source: http://news.xinhuanet.com/local/2013-03/02/c_124406508.htm, last accessed on 29 March 2016.

⁴ <Further notification on helping the special difficult families in the family planning>, issued by The National Health and Family Planning Commission, Ministry of Civil Affairs, Ministry of Finance, Ministry of Human Resources and Social Security, People’s Republic of China on 18 December 2013. Source:http://www.chinapop.gov.cn/jtfzs/s3582/201312/a8b82d3f888b4ffbaaf9486a64fe4121.shtml, last access on 29 March 2016.

⁵ http://www.gdwst.gov.cn/a/zcwj/2014122912746.html, last accessed on 7 August 2016.

⁶ http://www.sxsjsx.com/Announcements/tongzhigonggao/news_63374.html, last accessed on 7 August 2016.

⁷ Tibet was excluded from the study. Two other provinces, Hainan and Ningxia, both very small in population size, are not represented among the CHARLS counties.

⁸ If one member of a household aged 45 or over is sampled, then his or her spouse is also included without regard to the spouse’s age.

⁹ Potential measurement errors could bias our results. Some parents who experienced the death of their only child or all of their children may be likely to refuse to report that they had children, and will be excluded from our sample. However, the proportion of such misreports should be minor and should not largely affect our results. Still, we address this concern by including the observations with no children. Those reporting losing all their children on average have worse mental health than those reporting no birth at all (see Table A1 in the Appendix). Thus, if some bereaved parents are mislabelled as the no-children group, the measurement error will underestimate the gap in mental well-being between these two groups, i.e. underestimate the negative impact of losing all children, compared with the group with no children. Thus, our estimate may serve as the lower bound of the adverse impact of losing children.

¹⁰ The detailed coding is as follows. (1) no formal education (illiterate); (2) did not finish primary school but capable of reading and/or writing; (3) sishu/home school; (4) elementary school; (5) middle school; (6) high school; (7) vocational school; (8) 2-/3-year college/associate degree; (9) 4-year college/bachelor’s degree; (10) master’s degree; (11) doctoral degree/Ph.D.

¹¹ Unfortunately, subject to the questionnaire, we cannot further divide the group ‘others’ into unemployed, out of labour force, etc.

¹² For the self-reported subjective health status, many observations have missing values and thus we adopt the chronic disease dummy, rather than subjective health, as the health indicator.

¹³ Our final sample covers 448 rural villages or urban communities.

¹⁴ To save space, we only report the results with full controls for the regressions on happiness and loneliness. Results with partial controls are available upon request.

¹⁵ No gender heterogeneity is found in results on happiness and loneliness either. To save space, we only report the results on mental health. Results on happiness and loneliness are available upon request.

¹⁶ We have the birth year and death year for most of the dead children (669 out of 830 cases). We drop one observation because of the negative age. The average age at which the child died is 17.8.

¹⁷ There exists two types of omitted variable biases that limit the causal inference in our study. First, the bearing and birth of children may correlate with other unobserved variables that also affect mental health, happiness and loneliness, and the decision to give birth may be largely endogenous. Second, those losing children may have disadvantaged family backgrounds (including family health endowments), which also affect mental health. We address this concern partly in our analysis by controlling the parents’ health and income, but this is not a complete set of family background. We leave the further causal inference for future studies.

Additional information

Funding

This work was supported by the Youth Grant for Humanities and Social Sciences Research, Chinese Ministry of Education [grant number 14YJC790156] and by the Fundamental Research Funds for the Central Universities [grant number 20720161075].